Download The Human Gene AHNAK Encodes a Large

The Human Gene AHNAK Encodes a Large
Phosphoprotein Located Primarily in the Nucleus
E m m a Shtivelman* a n d J. Michael Bishop**
* Departments of Microbiology and Immunology, Biochemistry and Biophysics; and *The G. W. Hooper Research Foundation and
University of California, San Francisco, California 94143
Abstract. AHNAK is a newly identified human gene
notable for the exceptional size (c.a. 700 kD) and
structure of its product, and for the repression of its
expression in human neuroblastoma cells. Here we report the identification and partial characterization of
the protein encoded by AHNAK. The protein is located principally (but not exclusively) in the nucleus
and is phosphorylated on both serine and threonine.
The abundance of the protein increases appreciably
when cells withdraw from the division cycle, in response to either withdrawal of serum (fibroblasts) or
differentiation (neuroblastorna cells). By contrast, the
amount of phosphorylation appears to diminish in
those settings. The considerable abundance and conjectured fibrous structure of AHNAK protein suggest a
role in cytoarchitecture, but no function can yet be
discerned.
E have described previously the identification of a
human gene, AHNAK, that appears to encode a
protein of exceptional size (c.a. 700 kD) and structure (7). AHNAK was first encountered as one of a group
of eDNA clones isolated through subtractive cloning on the
basis of differential expression in human tumors of neuroectodermal origin (6). Another eDNA clone from the subtraeted eDNA library was identified as CD44, the integral
membrane glyeoprotein implicated in a variety of specific
cell-cell and cell-matrix interactions (6). Both AHNAK and
CD44 are expressed in melanoma and pheochromocytoma but
are repressed in neuroblastoma cell lines.
Our interest in AHNAK was prompted by the following
observations. First, we have documented a pattern of coordinated expression of AHNAK and CIM4 in different human
cell lineages (6, 7). This may indicate that both genes are
subject to the same transcriptional regulation, leading to the
suppression of their expression in neuroblastoma cell lines.
Analysis of this regulation might reveal factors involved in
the pathogenesis of neuroblastoma. Second, our analysis of
AHNAK revealed that its predicted protein product has an
unprecedented secondary structure, suggesting that it may
fulfill a novel function in the cell. The possible importance
of this function is reflected in the apparent phylogenetic conservation of AHNAK sequences (7).
AHNAK is expressed as a 17.5-kb RNA in a variety of
cells. Among human cell lines examined, we have found diminished or undetectable levels of expression in neuroblastomas, small cell lung carcinomas, and Burkitt lymphomas
(7). AHNAK contains an open reading frame that could be
translated into a 700-kD protein product. The salient feature
of the predicted protein product is the presence of repeated
domains comprising almost 80% of its amino acid sequence.
Approximately 4,300 amino acids in the middle portion of
the protein are arranged in multiple repeats of a sequence
motif of 128 amino acids. Some of the repeats have internal
deletions which occur at the same position. The degree of
amino acid identity between any two of the repeats is at least
80% and practically all the sequence differences are conservative (7).
Examination of amino acid sequence within the repeated
domain has revealed redundancy at a different level: heptad
repeats with a proline residue in the seventh position. The
character of each heptad could be best represented as
~ + ~ P + ~ + where ~ is a hydrophobic residue, + is a
charged residue, and P is proline. We proposed that AHNAK
protein has an unprecedented secondary structure resembling a/3-strand and distinct in that it has a periodicity of
2.33, i.e., three turns of the carbon backbone every seven
residues (7). This structure segregates the hydrophllic and
hydrophobic residues to opposing sides of the polypeptide
chain. We have also proposed that seven or eight of these
molecules could interact through their hydrophobic surfaces
to form an extremely long (,~1.2 t~) and thin (9.8 or 10A*)
barrel (7). Thus, AHNAK appears to represent a previously
unidentified type of fibrous protein. As a first step towards
discerning the function of AHNAK we have developed
specific antibodies for the AHNAK protein, and then used
these to show that the protein resides in the nucleus and is
pbospborylated on serine and threonine, particularly in
proliferating cells.
W
Dr. Shtivelman's current address is Systernix, 3400 West Bayshore Road,
Palo Alto, CA 94303.
9 The Rockefeller University Press, 0021-9525/93/02/625/6 $2.00
The Journal of Cell Biology, Volume 120, Number 3, February 1993 625-630
625
Materials and Methods
Preparation of Antisera
The chemically synthesized peptides CKISMPDVDLHLKGPKM (KIS)
and CLKMPKVKMPKFSMPGF (FEN) were used to raise specific antiAHNAK sera. An additional cysteine group was added to the arninotermini
to allow coupling of the peptides to the tuberculin purified protein derivative
with m-maieidobenzoyl-N-hydroxysuccinimide ether as described (3).
Each preparation was injected into two New Zealand rabbits under the
auspices of CaiTag, Inc. (South San Francisco, CA).
Immunoaltinity purification of antisera was performed on columns containing immunogenic peptide coupled to cyanogen-bromide activated
agarose beads as described (3). The antibodies eluted from the column were
dialyzed against PBS and used for protein analysis without further concentration. All results reported here were obtained using these preparations of
antibodies, which hereafter are named KIS4 and FEN2, and are derived
from sera obtained from rabbits injected with the KIS and FEN peptides,
respectively.
lmmunoprecipitation and lmmunobiot Analysis
In vivo labeling of cells for immunoprecipitation analysis was performed
bY35incubation of cells in methionine-free medium containing 250 ttCi of
[3 S]methionine per ml or in pbospbate-free medium containing 500 ttCi
[32p]orthophosphate per mi. Protein extracts were prepared from cultured
cells by lysis in RIPA buffer (150 mM NaC1, 20 mM Tris, pH 8.0, 5 mM
EDTA, 0.5% NP-40, 0.5% deoxycholate, 0.5% Tween, 0.25% SDS) or by
lysis in a buffer containing 150 mM NaCI, 20 mM Tris, pH 8.0, 5 mM
EDTA, and 0.5% of NP-40. All buffers contained the following protease inhibitors: PMSF at 1 mM, aprotinin at 50 ~tg/ml, leupeptin, and pepstatin
at 10 ttg/mi each. The lysates were clarified by centrifugation. Immunoprecipitations were carried out in RIPA buffer with 10 to 30 t~l of the
immunoaffinity purified antiserum KIS4 for 2 h at 40C. After the addition
of protein A-Sepharose, incubation was continued for another 40 min and
the beads were washed four times in RIPA buffer. The immunoprecipitated
proteins were analyzed by gel electrophoresis and autoradiography.
Cellular lysates for Western blotting were prepared by boiling pelleted
cells in Laemmli sample buffer. Proteins separated on 4% SDS-polyacrylamide gel with 3 % upper gel were transferred to nitrocellulose membranes
in a semi-dry blotting apparatus (E & K Scientific Products, Inc., Saratoga,
CA). The blots were incubated 30 rain in a solution containing 150 mM
NaCI, 5 mM KC1, 10 raM Tris-HC1, pH 7.0, 2% dry milk, 5% calf serum,
and 0.5 % Tween. The antibody incubation was for 2 h in the same solution.
The antigen-antibody complexes were detected with ~25I-containing antibody against rabbit IgG.
Subcellular Fractionation
Approximately 107 HeLa cells were washed in hypotonic buffer containing
10 mM KCI, 10 mM Tris HCI, pH 7.4, I mM MgC12, and protease inhibitors, and suspended in the same buffer containing 0.5% NP-40. After a 10min incubation on ice, nuclei were separated from the cytoplasmic fraction
by centrifugation at 1,000 g for 5 rain, washed in hypotonic buffer with 0.5 %
NP-40 and extracted in a buffer containing 150 mM NaCI. The wash fractions were combined with the cytoplasmic supernatant. Total cell lysate, nuclear, and cytoplasmic fractions were analyzed by Western blotting with
anti-AHNAK antibodies.
Immunofluorescence
Cells were grown on 12-mm coverslips for at least 24 h before processing.
After two washes in PBS, cells were fixed in 3.75% paraformaldehyde for
20 min at room temperature or in methanol for 6 min at -200C. Both fixation methods produced identical patterns of immunofluorescence. Fixed
cells were washed three times with PBS, permeabilized in Ab buffer (PBS,
3% BSA, 0.1% Triton X-100) for 20 rain, and incubated with the immunoaffinlty purified serum FEN2 for 1 h. The purified antibody and whole
serum FEN2 produced identical immunofluorescence patterns in a variety
of cells. Coverslips were washed with PBS and incubated for 40 min with
the secondary antibodies (affinity-purified rhodamine-conjugated donkey
anti-rabbit or fluorescein-conjugated sheep anti-rabbit (BoehringerMannheim Biochemieals, Indianapolis, IN) diluted 1:100 in Ab buffer.
Staining for DNA was done by incubation with Hoechst 33258 at 5/.tg/ml
in PBS. Coverslips were washed in PBS and mounted in Immumount (Shan-
The Journal of Cell Biology. Volume 120, 1993
don Southern Instruments Inc., Sewickley, PA). Micrographs were taken
with a Zeiss Axiophot microscope (Carl Zeiss, Oberkochen, Germany)
with the X63 oil objective and Kodak T-Max 400 film (Eastman Kodak Co.,
Rochester, NY).
Phosphoamino Acid Analysis
The AHNAK protein was immunoprecipitated from cells labeled with
[32p]orthophosphate, and then subjected to electrophoresis. The protein
was transferred to Immobilon-P (Millipore Continental Water Systems,
Bedford, MA), the labeled band was excised, and then hydrolyzed in 6 M
HC1 at ll0*C for 90 rain. Phospboamino acids were separated on cellulose
thin layer plates by electrophoresis in two dimensions (2). The migration
of unlabeled phosphoamino acids was determined by ninhydrin staining,
and radioactive amino acids were detected by autoradiography.
Results
Identification of AHNAK Protein
To identify and characterize the protein product of the
AHNAK gene, we have raised anti-AHNAK polyclonal sera
in rabbits. Based on the predicted amino acid sequence of
AHNAK protein, two synthetic peptides of 17 amino acids
each were made and used for immunizations. Both peptides
corresponded to regions within the repeated domain of the
predicted AHNAK polypeptide. The peptide "KIS" corresponds to the first 16 amino acids of the AHNAK 128 amino
acids repeat (7), while the peptide "FEN" corresponds to
amino acids 59-74 of the repeat.
The initial characterization of the sera was done using portions of AHNAK protein expressed in Escherichia coli. The
cDNA fragments of AHNAK were cloned into appropriate
pGEX vectors (11) to produce fusion proteins of glutathionS-transferase (GST) 1 and AHNAK-specific polypeptides.
Extracts from bacterial clones induced to express the fusion
proteins were analyzed by Western blotting with the antisera
raised against AHNAK peptides. We have found that sera
from one rabbit injected with the "KIS" peptide (designated
KIS4) and one injected with the "FEN-" peptide (FEN2) produced antibodies recognizing appropriate GST-AHNAK fusion proteins (data not shown). These two antisera were further examined by Western blot analysis of total protein
extracts prepared from cells expressing or lacking AHNAK
RNA. Fig. 1 A shows a Coomassie-stained gel containing
electrophoretically separated high molecular weight proteins
from the melanoma cell line C32r (expressing AHNAK
RNA) and small cell lung carcinoma line N417 (lacking
AHNAK RNA) (7). A very slowly migrating protein species, visible at the top of the gel in the C32r extract but not
in the N417 extract, appears to be recognized by various
anti-AHNAK antisera (FEN2 and KIS4 and the immunoaffinity-purified antibodies from these sera), but not the
preimmune sera (Fig. 1, B-F).
Fig. 2 shows that the candidate AHNAK protein was
specifically immunoprecipitated from extracts of psS]methionine-labeled cells with the immunoafllnity-purified antibody KIS4. No protein could be immunoprecipitated from
cell line N417, which lacks detectable levels of AHNAK
RNA, or when preimmune serum was used. The interaction
of antibodies with the protein was blocked by addition of the
immunogenic peptide. The molecular weight of AHNAK
protein could not be estimated on the basis of its migration
1. Abbreviation used in this paper: GST, glutathion-S-transferase.
626
Figure I. Western blot analysis of AHNAK protein with
polyclonal immune sera. Total protein extracts (150 #g)
from the melanoma cell line
C32r (lanes/) and small ceil
lung carcinoma cell line N417
(lanes 2) were separated on a
4% acrylamide gel.(.4) Coomassie blue stain gel. (B-G)
Identical Western blot strips
containing proteins shown in
A probed with: preimmune serum KIS4 diluted 1:200 (B);
immune serum KIS4 diluted
1:200 (C); immunoaffinitypurified antibodies from the
serum KIS4 diluted 1:100 (D);
preimmune serum FEN2 diluted 1:100 (E); immune serum FEN2 diluted 1:100 (F);
and immunoaffinity-purified
antibodies from serum FEN2
diluted 1:10 (G).
in the gel: the protein band appeared much higher on the gel
than the largest molecular weight marker of 200 kD. Our estimate of the molecular weight of AHNAK protein, based on
the length of the open reading frame of AHNAK RNA, is
~700 kD (7).
In further characterization of the antisera, we found that
KIS4 recognized only the denatured form of AHNAK protein and gave no signal in immunofluorescence. In contrast,
FEN2 reacted with both native and denatured protein, and
could be used in immunofluorescence. Since immunoaffinity-purified KIS4 had a significantly higher titer than FEN2,
we used KIS4 for all analyses except immunofluorescence.
All immunoaflinity-purified preparations of the two antisera
appeared to have the same specificity when tested accordingly (see above).
The antibodies raised against AHNAK protein were used to
address the question of its subcellular localization. For convenlence we analyzed HeLa cells, which are readily fractionated into subcellular organelles and which produce
abundant AHNAK protein (see below). HeLa cells were
fractionated into crude nuclear and cytoplasmic fractions by
lysis in hypotonic buffer in the presence of nonionic detergent. Proteins were extracted from nuclei by 150 mM NaC1
in the presence of 0.5% NP-40. Fig. 3 shows Western blot
analysis of AHNAK in the nuclear and cytoplasmic fractions
of HeLa cells. The major portion of AHNAK protein was detected in the nuclear rather than the cytoplasmic fraction.
Next, we used indirect immunofluorescence to examine
the subcellular localization of AHNAK. Fig. 4 shows the
Figure 2. Immunoprecipitation of AHNAK protein. Cells
were labeled with [~SS]methionine and lysed in RIPA buffer. Equal amounts of protein
extracts (200 ~g) were analyzed by immunoprecipitation
with 20 ~1of the immunoaffinity-purified antibodies KIS4.
The following ceil lines were
analyzed: melanoma C32r (/);
small ceil lung carcinomaN417
(2); neuroblastoma NGP (3);
mouse fibroblasts NIH 3T3
(4); NIH 3T3 with addition of
2/~g of the immunogenic peptide KIS (5); and NIH Yr3
extract immunoprecipitated
with the preimmune serum (6).
Figure 3. SubcellularlocaliTationof AHNAK
protein in HeLa ceils. Subceilular fractions
were prepared as described in Materials and
Methods. Proteins extracted from the equivalent of 10~cells were analyzed by Western
blotting with the immunoaffinity-purified
antiserum KIS4. T, total cell lysate; N, nuclear fraction; and C, cytoplasmic fraction.
Shtivelmanand BishopAHNAKProtein
SubceUular Localization of A H N A K Protein
627
Figure 4. Localization of
AHNAK protein in cultured
cells by indirect immunofluorescence. NIH 3T3 cells (a-f)
or Ptkl (g and h) cells were
stained with Hcechst 33258
(a, c, e, and g), preimmune
serum (b), anti-AHNAK immunoaffinity-purified serum
FEN2 in presence (d), or absence (land h) of the immune
peptide followedby secondary
antibody conjugatedto rhodamine (b, d, and f) or fluorescein (h).
pattern of staining observed with NIH 3"1"3cells. Specific
staining of proteins in the nuclei and in an asymmetric
perinuclear structure was evident with the immune antibody,
but not with the preimmune serum or when the antibodies
were preincubated with the immunogenic peptide (Fig. 4).
In addition, we consistently observed a weak, diffuse cytoplasmic staining. The predominantly nuclear localization of
AHNAK protein by immunofluorescence is in agreement
with the results of the cell fractionations.
The pednuclear structure stained by the antibodies resembled morphologically the Golgi apparatus. To examine this
possibility we performed double staining of cells with FEN2
and wheat germ agglutinin, which predominantly binds to
the elements of the Golgi network. We observed that the
wheat germ agglutinin and the anti-AHNAK antibody bound
to the same perinuclear structure. Thus, it remains possible
that AHNAK protein might be associated with the Golgi network (data not shown), but that possibility requires further
assessment (see Discussion).
We have performed immunofluorescent localization of
AHNAK protein in several human cells lines, including epitheliai cells, melanoma and neuroblastoma cells. Independent of the cell line we have observed staining of nuclei and
a perinuclear asymmetric structure resembling the Golgi
network (data not shown). We have also examined the rat
kangaroo kidney cell line Ptkt for the localization of AHNAK
protein after confirming that the same high molecular weight
protein could be immunoprecipitated from extracts of these
The Journal of Cell Biology, Volume 120, 1993
628
Figure5. Analysisof AHNAK
protein in growing versus
quiescent NIH 3"1"3cells. (A) 2
x 105 cells from actively
proliferating (lane/) or quiescent (lane 2) culture were labeled with inorganic [32p] for
2.5 h, lysed in RIPA buffer,
and analyzed by immunoprecipitation with the immunoaffinity-purified KIS4 antibodies. The smear in lane I is due
to the presence of radioactive
DNA in the cell lysate. (B)
l0 s cells from parallel cultures were analyzed by Western blotting with antibodies
KIS4.
cells (data not shown). Again, we have observed immunofluorescence in the nuclei and the structure resembling the
Golgi apparatus (Fig. 4 H).
A H N A K Is a Phosphoprotein
We have reported previously that the carboxyl-terminal portion of the AHNAK sequence contains several repeats of the
sequence "lysine-serine-proline-lysine" which is identical to
the "DNA finger" sequences found in historic proteins (4).
The "DNA fingers" found in histories are the sites of mitosisspecific phosphorylation on serine residues (4). We conducted experiments to determine if AHNAK protein is phosphorylated and, if so, whether the modification is related to
the cell cycle. To this end, AHNAK protein was immunoprecipitated from NIH 3T3 cells after metabolic labeling with
radioactive inorganic phosphate. We examined actively
growing and quiescent cells deprived of serum for 2 d. Fig.
5 A shows that AHNAK protein was phosphorylated in
growing cells and that this phosphorylation was greatly diminished in quiescent NIH 3"173cells. In contrast, Western
blot analysis demonstrated that quiescent cells contain
higher levels of AHNAK protein (Fig. 5 B). We conclude
that AHNAK is a phosphoprotein and that the transition of
fibroblasts into Go is accompanied by both accumulation
and apparent dephosphorylation of AHNAK protein.
We have performed phosphoamino acid analysis of
AHNAK protein in actively growing NIH 3T3 ceUs and have
determined that it is phosphorylated on serine and, to a much
lower extent, on threonine (data not shown). Serine residues
are found in four conserved positions within every repeat
unit of AHNAK (7) with occasional substitutions by threonine
residues. The carboxy-terminal unique portion of AHNAK
is relatively serine rich (7), while threonine residues are infrequent. Tyrosine residues are found at a conserved position
in some of the repeats of AHNAK, but no phosphorylation
of tyrosine was observed (data not shown). Identical results
were obtained when phosphoamino acid analysis was performed with AHNAK protein immunoprecipitated from
HeLa cells (data not shown).
lqgure 6. Effect of differentiationon expression of AHNAK RNA.
Cells were treated for the periods of time indicated and total cellular RNAs (20 #g) were analyzed for the levels of AHNAK RNA
by RNAse protection assay of a specific riboprobe prepared with
a 5' end fragment of AHNAK cDNA (7). The length of the protected fragment is 330 nucleotides. The lane labeled Mel contains
20 #g of total RNA from the melanoma cell line HT144.
reduced in neuroblastoma cell lines in comparison to other,
more differentiated cells of neuroectodermal origin (6, 7).
We examined if any changes in the expression of AHNAK
occur in differentiating neuroblastoma cells. Neuroblastoma
cell lines NGP and LAN5 were treated with retinoic acid to
induce neuronal differentiation (8) and analyzed for the levels of AHNAK RNA over the course of treatment. In both
cell lines we observed an increase in the levels of AHNAK
RNA after 1-2 d of exposure to retinoic acid, after which the
levels of AHNAK RNA remained unchanged (Fig. 6). The
long interval between application of retinoic acid and the induction of AHNAK RNA most likely indicates that the latter
is an indirect consequence of more immediate changes induced by retinoic acid. The timing of AHNAK RNA induction in neuroblastoma cells coincides with the cessation of
cellular division and the appearance of morphological
changes in the transformed neuroblasts (8, 9).
Next we analyzed the levels and phosphorylation of
AHNAK protein in untreated versus differentiating neuroblastoma cells (Fig. 7). AHNAK protein labeled with radioactive phosphate was precipitated from untreated cells of
neuroblastoma NGP and after treatment with retinoic acid
for 2 d. Fig. 7 (A and B) shows that the phosphate content
of AHNAK protein in differentiating cells was diminished,
whereas the levels of the protein increased. Similar results
were obtained with the neuroblastoma cell line LAN5 (data
not shown). The levels of AHNAK remained elevated for up
to 7 d after the beginning of the treatment (Fig. 7 C).
Discussion
The Size of A H N A K Protein
AHNAK was first identified as a gene whose expression is
As predicted from nucleotide sequence (7), AHNAK protein
isolated from cells appears to be exceptionally large. The
data presented here do not permit a reliable assessment of
size because large proteins may migrate anomalously in
Shtivelmanand BishopAHNAKProtein
629
A H N A K Protein in Differentiating
Neuroblastoma Cells
Figure 7. Effect of differentiation on expression and phosphorylation of AHNAK protein. (.4) NGP cells were
treated with 5 izM of retinoic
acid for 48 h. 2 x 105 cells
from untreated (lane NGP)
and treated with retinoic acid
culture (lane NGP + RA)
were labeled with inorganic
[32p] for 2.5 h, lysed in RIPA
buffer, and analyzedby immunoprecipitation with the immunoaffinity purified serum
KIS4. (B) 2 • lit cells from
parallel cultures were analyzed by Westernblotting with
antibodies KIS4. (C) Western
blot analysis of protein lysates
(100 t~g of protein per lane)
prepared from NGP ceils
treated with retinoic acid for
the periods of time indicated.
polyacrylamide gels. Thus, we continue to use the estimate
of 700 kD for the mass of the protein, based on the deduced
amino acid sequence. The representation of AHNAK in
cDNAs still suffers from two apparently small gaps, but the
corresponding regions of the genomic DNA have been sequenced and found to contain open reading frames that are
continuous with the open reading frame represented in the
adjoining cDNAs (7). It seems likely that we have the entire
coding domain for AHNAK in view. We consider it unlikely
that the gene produces a second protein by means of alternative splicing, since no introns are apparent in the nucleotide
sequence of AHNAK (7). Therefore, we conclude that the
protein identified here with two distinctive antisera and in diverse cell lines is the principal or sole product of AHNAK.
The SubceUular Location of AHNAK Protein
Our results indicate that the bulk of AHNAK protein is located in the nucleus. How can a protein with a mass of 700
kD gain access to the nucleus? Two properties of the AHNAK
proteins may provide an explanation. First, the rodlike structure conjectured for the protein may facilitate passage
through the confining diameter of nuclear pores. Second, the
carboxy-terminal domain of the protein contains several possible nuclear localization signals that might facilitate nuclear
transport by increasing the efficiency of interaction with proteins that bind the nuclear localization signals (10).
Several features of the AHNAK protein suggest that it may
have a role in the architecture of the nucleus. First, the protein is both relatively stable (data not shown) and abundant
(Fig. 1 A). Second, molecular modeling predicts a remarkably elongated and rigid tertiary structure (7). And third, our
preliminary results indicate that the protein might bind DNA
in a nonspecific manner (data not shown). These properties
cause one to think of the nuclear matrix or scaffolding (1, 5),
but the AHNAK protein is too readily extractable from the
nucleus to be considered a component of these structures. Instead, the decidedly hydrophilic surface of the projected
structure suggests that the protein could easily reside in the
nucleoplasm. Moreover, it is possible that the two terminal
domains of the protein (which are not part of the fibrous
structure) could have functions of their own that are not of
an architectural nature.
The nuclear staining of AHNAK protein was mainly punctate, an observation that seems at odds with the predicted
structure for the protein. Given the denaturation that accompanies fixation, however, it seems unreasonable to expect the
pattern of staining to necessarily reflect the native structure
of the protein. Examination by immunoelectron microscopy
might be more revealing.
Analysis by immunofluorescence indicated that a minor
fraction of AHNAK protein may be associated with the
Golgi network. The protein displays no structural feature
that would either predict or explain sequestration in a membranous organelle. Our preliminary data from subcellular
fractionations confirm that the cytoplasmic AHNAK protein
is confined to the membrane-containing fraction, but further
work will be required to determine whether the protein does
in reality reside in the Golgi apparatus.
Modifications of the A H N A K Protein During Changes
in Cellular Proliferation and Differentiation
The product of AHNAK is phosphorylated on serine and
threonine. Both amino acids are abundant throughout the
protein, but the particular residues subject to phosphorylation and the responsible kinase(s) are not known. The phosphorylation appears to decrease substantially when ceils are
rendered quiescent. It is conceivable that this decrease is due
to limitations of isotopic labeling in quiescent cells, but we
note that the decrease was observed with cells rendered
quiescent by either deprivation of serum or induction of
differentiation in rich growth medium. By contrast, the
abundance of AHNAK protein increases appreciably in the
Go stage of the cell cycle, as if the protein might have an especially prominent function in quiescent cells. Perhaps dephosphorylation is a correlate of that function.
We are grateful to Gary Ramsay and Nancy Quintreli for advice on immunological techniques.
Work summarized here was supported by the National Institutes of
Health grant No. CA 44338 and by funds from the G. W. Hoopcr Research
Foundation.
Received for publication 25 June 1992 and in revised form 15 September
1992.
References
1. Bereaney, R., and D. S. Coffey. 1974. Identification of a nuclear protein
matrix. Biochem. Biophys. Res. Commun. 60:1410-1417.
2. Cooper, J. A., B. Sefton, andT. Hunter. 1983. Detection and quantification
of phosphotyrosine in proteins. Methods Enzymol. 99:387--402.
3. Harlow, E., and D. Lane. 1988. Antibodies. A Laboratory Manual. Cold
Spring Harbor Laboratory, Cold Spring Harbor, NY. 726 pp.
4. Langan, T. A., J. Gautier, M. Lohka, R. Hollingsworth, S. Moreno, P.
Nurse, J. Mailer, and R. A. Selafani. 1989. Mammalian growthassociated H1 historic kinase: a homolog of edc2§
protein kinases controlling mitotic entry in yeast and frog cells. Mol. Cell. Biol.
9:3860-3868.
5. Mirkovitch, J., M. E. Mirault, and U. K. Laemmli. 1984. Organization of
the higher-order chromatin loop: specific DNA attachment sites on nuclear scaffold. Cell. 39:223-232.
6. Shtivehnan, E., and J. M. Bishop. 1991. Expression of CD44 is repressed
in neuroblastoma cells. Idol. Cell. Biol. 11:5446-5453.
7. Shtivelrnan, E., F. E. Cohen, and J. M. Bishop. 1992. A human gene
(AHNAK) encoding an unusually large protein with an unprecedented
structure. Proc. Natl. Acad. Sci. USA. 89:5472-5476.
8. Sidell, N. 1982. Retinoic acid-induced growth inhibition and morphological differentiation of human neuroblastuma cells in vitro. J. Natl. Cancer
Inst. 68:589-593.
9. Sidell, N., A. Altman, M. R. Haussler, and R. C. Sr162 1983. Effvcts
of retinoic acid on the growth and phenotypic expression of several human neuroblastoma cell lines. J. Neuroscl. Res. 8:375-391.
10. Silver, P. A. 1991. How proteins enter the nucleus. Cell. 64:489--497.
11. Smith, D. B., and K. S. Johnson. 1988. Single-step purification of polypeptides expressed in Escherichia coil as fusions with glutathione
S-transferase. Gene. 67:31-40.